1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
|
/**
* Copyright (C) 2020-present MongoDB, Inc.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the Server Side Public License, version 1,
* as published by MongoDB, Inc.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* Server Side Public License for more details.
*
* You should have received a copy of the Server Side Public License
* along with this program. If not, see
* <http://www.mongodb.com/licensing/server-side-public-license>.
*
* As a special exception, the copyright holders give permission to link the
* code of portions of this program with the OpenSSL library under certain
* conditions as described in each individual source file and distribute
* linked combinations including the program with the OpenSSL library. You
* must comply with the Server Side Public License in all respects for
* all of the code used other than as permitted herein. If you modify file(s)
* with this exception, you may extend this exception to your version of the
* file(s), but you are not obligated to do so. If you do not wish to do so,
* delete this exception statement from your version. If you delete this
* exception statement from all source files in the program, then also delete
* it in the license file.
*/
#include "mongo/platform/basic.h"
#include "mongo/db/catalog/catalog_test_fixture.h"
#include "mongo/db/catalog/create_collection.h"
#include "mongo/db/timeseries/bucket_catalog.h"
#include "mongo/db/views/view_catalog.h"
#include "mongo/stdx/future.h"
#include "mongo/unittest/bson_test_util.h"
#include "mongo/unittest/death_test.h"
#include "mongo/util/concurrency/thread_pool.h"
namespace mongo {
namespace {
class BucketCatalogTest : public CatalogTestFixture {
protected:
class Task {
AtomicWord<bool> _running{false};
stdx::packaged_task<void()> _task;
stdx::future<void> _future;
stdx::thread _taskThread;
public:
Task(std::function<void()>&& fn);
~Task();
const stdx::future<void>& future();
};
void setUp() override;
virtual BSONObj _makeTimeseriesOptionsForCreate() const;
void _commit(const std::shared_ptr<BucketCatalog::WriteBatch>& batch,
uint16_t numPreviouslyCommittedMeasurements,
size_t expectedBatchSize = 1);
void _insertOneAndCommit(const NamespaceString& ns,
uint16_t numPreviouslyCommittedMeasurements);
long long _getNumWaits(const NamespaceString& ns);
OperationContext* _opCtx;
BucketCatalog* _bucketCatalog;
StringData _timeField = "time";
StringData _metaField = "meta";
NamespaceString _ns1{"bucket_catalog_test_1", "t_1"};
NamespaceString _ns2{"bucket_catalog_test_1", "t_2"};
NamespaceString _ns3{"bucket_catalog_test_2", "t_1"};
BucketCatalog::CommitInfo _commitInfo{StatusWith<SingleWriteResult>(SingleWriteResult{})};
};
class BucketCatalogWithoutMetadataTest : public BucketCatalogTest {
protected:
BSONObj _makeTimeseriesOptionsForCreate() const override;
};
BucketCatalogTest::Task::Task(std::function<void()>&& fn)
: _task{[this, fn = std::move(fn)]() {
_running.store(true);
fn();
}},
_future{_task.get_future()},
_taskThread{std::move(_task)} {
while (!_running.load()) {
stdx::this_thread::yield();
}
}
BucketCatalogTest::Task::~Task() {
_taskThread.join();
}
const stdx::future<void>& BucketCatalogTest::Task::future() {
return _future;
}
void BucketCatalogTest::setUp() {
CatalogTestFixture::setUp();
_opCtx = operationContext();
_bucketCatalog = &BucketCatalog::get(_opCtx);
for (const auto& ns : {_ns1, _ns2, _ns3}) {
ASSERT_OK(createCollection(
_opCtx,
ns.db().toString(),
BSON("create" << ns.coll() << "timeseries" << _makeTimeseriesOptionsForCreate())));
}
}
BSONObj BucketCatalogTest::_makeTimeseriesOptionsForCreate() const {
return BSON("timeField" << _timeField << "metaField" << _metaField);
}
BSONObj BucketCatalogWithoutMetadataTest::_makeTimeseriesOptionsForCreate() const {
return BSON("timeField" << _timeField);
}
void BucketCatalogTest::_commit(const std::shared_ptr<BucketCatalog::WriteBatch>& batch,
uint16_t numPreviouslyCommittedMeasurements,
size_t expectedBatchSize) {
ASSERT(batch->claimCommitRights());
_bucketCatalog->prepareCommit(batch);
ASSERT_EQ(batch->measurements().size(), expectedBatchSize);
ASSERT_EQ(batch->numPreviouslyCommittedMeasurements(), numPreviouslyCommittedMeasurements);
_bucketCatalog->finish(batch, _commitInfo);
}
void BucketCatalogTest::_insertOneAndCommit(const NamespaceString& ns,
uint16_t numPreviouslyCommittedMeasurements) {
auto result = _bucketCatalog->insert(_opCtx,
ns,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
auto& batch = result.getValue();
_commit(batch, numPreviouslyCommittedMeasurements);
}
long long BucketCatalogTest::_getNumWaits(const NamespaceString& ns) {
BSONObjBuilder builder;
_bucketCatalog->appendExecutionStats(ns, &builder);
return builder.obj().getIntField("numWaits");
}
TEST_F(BucketCatalogTest, InsertIntoSameBucket) {
// The first insert should be able to take commit rights, but batch is still active
auto result1 =
_bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
auto batch1 = result1.getValue();
ASSERT(batch1->claimCommitRights());
ASSERT(batch1->active());
// A subsequent insert into the same bucket should land in the same batch, but not be able to
// claim commit rights
auto result2 =
_bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
auto batch2 = result2.getValue();
ASSERT_EQ(batch1, batch2);
ASSERT(!batch2->claimCommitRights());
// The batch hasn't actually been committed yet.
ASSERT(!batch1->finished());
_bucketCatalog->prepareCommit(batch1);
// Still not finished, but no longer active.
ASSERT(!batch1->finished());
ASSERT(!batch1->active());
// The batch should contain both documents since they belong in the same bucket and happened
// in the same commit epoch. Nothing else has been committed in this bucket yet.
ASSERT_EQ(batch1->measurements().size(), 2);
ASSERT_EQ(batch1->numPreviouslyCommittedMeasurements(), 0);
// Once the commit has occurred, the waiter should be notified.
_bucketCatalog->finish(batch1, _commitInfo);
ASSERT(batch2->finished());
auto result3 = batch2->getResult();
ASSERT_OK(result3.getStatus());
}
TEST_F(BucketCatalogTest, GetMetadataReturnsEmptyDocOnMissingBucket) {
auto batch = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow)
.getValue();
ASSERT(batch->claimCommitRights());
auto bucket = batch->bucket();
_bucketCatalog->abort(batch);
ASSERT_BSONOBJ_EQ(BSONObj(), _bucketCatalog->getMetadata(bucket));
}
TEST_F(BucketCatalogTest, InsertIntoDifferentBuckets) {
auto result1 =
_bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now() << _metaField << "123"),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
auto result2 =
_bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now() << _metaField << BSONObj()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
auto result3 =
_bucketCatalog->insert(_opCtx,
_ns2,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
// Inserts should all be into three distinct buckets (and therefore batches).
ASSERT_NE(result1.getValue(), result2.getValue());
ASSERT_NE(result1.getValue(), result3.getValue());
ASSERT_NE(result2.getValue(), result3.getValue());
// Check metadata in buckets.
ASSERT_BSONOBJ_EQ(BSON(_metaField << "123"),
_bucketCatalog->getMetadata(result1.getValue()->bucket()));
ASSERT_BSONOBJ_EQ(BSON(_metaField << BSONObj()),
_bucketCatalog->getMetadata(result2.getValue()->bucket()));
ASSERT_BSONOBJ_EQ(BSON(_metaField << BSONNULL),
_bucketCatalog->getMetadata(result3.getValue()->bucket()));
// Committing one bucket should only return the one document in that bucket and shoukd not
// affect the other bucket.
for (const auto& batch : {result1.getValue(), result2.getValue(), result3.getValue()}) {
_commit(batch, 0);
}
}
TEST_F(BucketCatalogTest, NumCommittedMeasurementsAccumulates) {
// The numCommittedMeasurements returned when committing should accumulate as more entries in
// the bucket are committed.
_insertOneAndCommit(_ns1, 0);
_insertOneAndCommit(_ns1, 1);
}
TEST_F(BucketCatalogTest, ClearNamespaceBuckets) {
_insertOneAndCommit(_ns1, 0);
_insertOneAndCommit(_ns2, 0);
_bucketCatalog->clear(_ns1);
_insertOneAndCommit(_ns1, 0);
_insertOneAndCommit(_ns2, 1);
}
TEST_F(BucketCatalogTest, ClearDatabaseBuckets) {
_insertOneAndCommit(_ns1, 0);
_insertOneAndCommit(_ns2, 0);
_insertOneAndCommit(_ns3, 0);
_bucketCatalog->clear(_ns1.db());
_insertOneAndCommit(_ns1, 0);
_insertOneAndCommit(_ns2, 0);
_insertOneAndCommit(_ns3, 1);
}
TEST_F(BucketCatalogTest, InsertBetweenPrepareAndFinish) {
auto batch1 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow)
.getValue();
ASSERT(batch1->claimCommitRights());
_bucketCatalog->prepareCommit(batch1);
ASSERT_EQ(batch1->measurements().size(), 1);
ASSERT_EQ(batch1->numPreviouslyCommittedMeasurements(), 0);
// Insert before finish so there's a second batch live at the same time.
auto batch2 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow)
.getValue();
ASSERT_NE(batch1, batch2);
_bucketCatalog->finish(batch1, _commitInfo);
ASSERT(batch1->finished());
// Verify the second batch still commits one doc, and that the first batch only commited one.
_commit(batch2, 1);
}
DEATH_TEST_F(BucketCatalogTest, CannotCommitWithoutRights, "invariant") {
auto result = _bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
auto& batch = result.getValue();
_bucketCatalog->prepareCommit(batch);
}
DEATH_TEST_F(BucketCatalogTest, CannotFinishUnpreparedBatch, "invariant") {
auto result = _bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
auto& batch = result.getValue();
ASSERT(batch->claimCommitRights());
_bucketCatalog->finish(batch, _commitInfo);
}
TEST_F(BucketCatalogWithoutMetadataTest, GetMetadataReturnsEmptyDoc) {
auto batch = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow)
.getValue();
ASSERT_BSONOBJ_EQ(BSONObj(), _bucketCatalog->getMetadata(batch->bucket()));
_commit(batch, 0);
}
TEST_F(BucketCatalogWithoutMetadataTest, CommitReturnsNewFields) {
// Creating a new bucket should return all fields from the initial measurement.
auto result = _bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now() << "a" << 0),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
ASSERT_OK(result);
auto batch = result.getValue();
auto oldId = batch->bucket()->id();
_commit(batch, 0);
ASSERT_EQ(2U, batch->newFieldNamesToBeInserted().size()) << batch->toBSON();
ASSERT(batch->newFieldNamesToBeInserted().count(_timeField)) << batch->toBSON();
ASSERT(batch->newFieldNamesToBeInserted().count("a")) << batch->toBSON();
// Inserting a new measurement with the same fields should return an empty set of new fields.
result = _bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now() << "a" << 1),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
ASSERT_OK(result);
batch = result.getValue();
_commit(batch, 1);
ASSERT_EQ(0U, batch->newFieldNamesToBeInserted().size()) << batch->toBSON();
// Insert a new measurement with the a new field.
result = _bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now() << "a" << 2 << "b" << 2),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
ASSERT_OK(result);
batch = result.getValue();
_commit(batch, 2);
ASSERT_EQ(1U, batch->newFieldNamesToBeInserted().size()) << batch->toBSON();
ASSERT(batch->newFieldNamesToBeInserted().count("b")) << batch->toBSON();
// Fill up the bucket.
for (auto i = 3; i < gTimeseriesBucketMaxCount; ++i) {
result = _bucketCatalog->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now() << "a" << i),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
ASSERT_OK(result);
batch = result.getValue();
_commit(batch, i);
ASSERT_EQ(0U, batch->newFieldNamesToBeInserted().size()) << i << ":" << batch->toBSON();
}
// When a bucket overflows, committing to the new overflow bucket should return the fields of
// the first measurement as new fields.
auto result2 = _bucketCatalog->insert(
_opCtx,
_ns1,
BSON(_timeField << Date_t::now() << "a" << gTimeseriesBucketMaxCount),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow);
auto& batch2 = result2.getValue();
ASSERT_NE(oldId, batch2->bucket()->id());
_commit(batch2, 0);
ASSERT_EQ(2U, batch2->newFieldNamesToBeInserted().size()) << batch2->toBSON();
ASSERT(batch2->newFieldNamesToBeInserted().count(_timeField)) << batch2->toBSON();
ASSERT(batch2->newFieldNamesToBeInserted().count("a")) << batch2->toBSON();
}
TEST_F(BucketCatalogTest, AbortBatchWithOutstandingInsertsOnBucket) {
auto batch1 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow)
.getValue();
ASSERT(batch1->claimCommitRights());
_bucketCatalog->prepareCommit(batch1);
ASSERT_EQ(batch1->measurements().size(), 1);
ASSERT_EQ(batch1->numPreviouslyCommittedMeasurements(), 0);
// Insert before finish so there's a second batch live at the same time.
auto batch2 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow)
.getValue();
ASSERT_NE(batch1, batch2);
ASSERT_EQ(0, _getNumWaits(_ns1));
// Aborting the batch will have to wait for the commit of batch1 to finish, then will proceed
// to abort batch2.
ASSERT(batch2->claimCommitRights());
auto task = Task{[&]() { _bucketCatalog->abort(batch2); }};
// Add a little extra wait to make sure abort actually gets to the blocking point.
stdx::this_thread::sleep_for(stdx::chrono::milliseconds(10));
ASSERT(task.future().valid());
ASSERT(stdx::future_status::timeout == task.future().wait_for(stdx::chrono::microseconds(1)))
<< "clear finished before expected";
_bucketCatalog->finish(batch1, _commitInfo);
ASSERT(batch1->finished());
// Now the clear should be able to continue, and will eventually abort batch2.
task.future().wait();
ASSERT_EQ(1, _getNumWaits(_ns1));
ASSERT(batch2->finished());
ASSERT_EQ(batch2->getResult().getStatus(), ErrorCodes::TimeseriesBucketCleared);
ASSERT_EQ(1, _getNumWaits(_ns1));
}
TEST_F(BucketCatalogTest, CombiningWithInsertsFromOtherClients) {
_opCtx->setLogicalSessionId(makeLogicalSessionIdForTest());
auto batch1 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kDisallow)
.getValue();
_opCtx->setLogicalSessionId(makeLogicalSessionIdForTest());
auto batch2 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kDisallow)
.getValue();
_opCtx->setLogicalSessionId(makeLogicalSessionIdForTest());
auto batch3 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow)
.getValue();
_opCtx->setLogicalSessionId(makeLogicalSessionIdForTest());
auto batch4 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kAllow)
.getValue();
ASSERT_NE(batch1, batch2);
ASSERT_NE(batch1, batch3);
ASSERT_NE(batch2, batch3);
ASSERT_EQ(batch3, batch4);
_commit(batch1, 0);
_commit(batch2, 1);
_commit(batch3, 2, 2);
}
TEST_F(BucketCatalogTest, CannotConcurrentlyCommitBatchesForSameBucket) {
_opCtx->setLogicalSessionId(makeLogicalSessionIdForTest());
auto batch1 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kDisallow)
.getValue();
_opCtx->setLogicalSessionId(makeLogicalSessionIdForTest());
auto batch2 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kDisallow)
.getValue();
ASSERT(batch1->claimCommitRights());
ASSERT(batch2->claimCommitRights());
// Batch 2 will not be able to commit until batch 1 has finished.
_bucketCatalog->prepareCommit(batch1);
auto task = Task{[&]() { _bucketCatalog->prepareCommit(batch2); }};
// Add a little extra wait to make sure prepareCommit actually gets to the blocking point.
stdx::this_thread::sleep_for(stdx::chrono::milliseconds(10));
ASSERT(task.future().valid());
ASSERT(stdx::future_status::timeout == task.future().wait_for(stdx::chrono::microseconds(1)))
<< "prepareCommit finished before expected";
_bucketCatalog->finish(batch1, _commitInfo);
task.future().wait();
_bucketCatalog->finish(batch2, _commitInfo);
}
TEST_F(BucketCatalogTest, DuplicateNewFieldNamesAcrossConcurrentBatches) {
_opCtx->setLogicalSessionId(makeLogicalSessionIdForTest());
auto batch1 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kDisallow)
.getValue();
_opCtx->setLogicalSessionId(makeLogicalSessionIdForTest());
auto batch2 = _bucketCatalog
->insert(_opCtx,
_ns1,
BSON(_timeField << Date_t::now()),
BucketCatalog::CombineWithInsertsFromOtherClients::kDisallow)
.getValue();
// Batch 2 is the first batch to commit the time field.
ASSERT(batch2->claimCommitRights());
_bucketCatalog->prepareCommit(batch2);
ASSERT_EQ(batch2->newFieldNamesToBeInserted().size(), 1);
ASSERT_EQ(*batch2->newFieldNamesToBeInserted().begin(), _timeField);
_bucketCatalog->finish(batch2, _commitInfo);
// Batch 1 was the first batch to insert the time field, but by commit time it was already
// committed by batch 2.
ASSERT(batch1->claimCommitRights());
_bucketCatalog->prepareCommit(batch1);
ASSERT(batch1->newFieldNamesToBeInserted().empty());
_bucketCatalog->finish(batch1, _commitInfo);
}
} // namespace
} // namespace mongo
|